Great stuff. I'm currently trying to build a REST API node backend, and have been frustrated with most tutorials giving overly basic examples that don't help with real-world application. This was comprehensive, crystal clear, and without fluff. You've probably saved me a week of banging my head on my keyboard, cheers!
I'm working on implementing Winston for a project, and found their "Quick Start" code confusing. Thank you so much for creating this video tutorial and breaking down what's happening even further. This was incredibly helpful!
This was helpful for someone who attempted to log their stuff for the first time. Though there could be 30 more minutes on when to log, what to log, how to log properly etc.
Noticed at 26:20 you needed to restart the app because you're making changes on the fly, is there a way to apply dynamic logging without restarting the app ??
So cool! Watching you from Ukraine. I love that you allways show how it has to be in real project and showing package structure. It really helps me on my first job!
How to daily rotate logs using Winston except the first day?? Like my initial log should be info.log then next day a new info.log should be created and previous one should be appended with a date
per default, the log is written to stdout so that it is just getting appended to the existing logs. This is mostly sufficient because you have everything in one place. Not sure if there is a use case for what you want to do because each log statement also contains a timestamp
Azure offers a dedicated log aggregator for this. Typically you do not need to configure anything, but just run a daemon in your cluster. The daemon will grab all output from std out and push it to the log aggregator
Awesome video, Kindly also make production logging with winston and elk, with also tracing the whole api request cycle, I have been looking for child loggers and tracing whole api request ?
Hi Shimron, assigning it a traceId would do the trick. This works especially well with a microservice-based system. Maybe check out opentelemetry.io/. If you only have 1 server / microservice, opentelemetry might be overkill though. It is rather meant for microservices
Wow, such an easy and wonderful video. Thanks for it. :) Can we make the logger work for any specific level, like just for error or warn, excluding the other higher levels? Please suggest something.
Hi Anshika, yes you can do that. If you only want to output warning or anything more severe, then you would need to set level: warning when you create the logger. So basically this level field means that it will log for warn logging level and below. I.e. you would log for warn and error. You can find a more detailed description here: github.com/winstonjs/winston#logging-levels
I've noticed when using with morgan that the morgan console log appears after winston. I'm assuming it's because morgan is async and winston is not? Do you know a good way to adjust winston logs to appear after morgan request logs? I think it looks nicer when you see the request and then all the details that follow in your logs.
Hi Nathan, I think it depends when the middleware is plugged in. If you plug it in right at the start then morgan logs should be the first things that show up: github.com/expressjs/morgan#expressconnect
Hi Arun, storing it in a file is probably the cheaper option. If you don't intend to search in the logs dynamically from some webapp, then I guess you can just run go with a file. Storing the logs in MongoDB (in a separate instance from your main db so it does not affect performance) would only make sense if you have a use case where you need to dynamically access the logs from your webapp and search in them and so on
i guess, logger.log('error',`error string`); logger.error('error string'); those commands do samething. if log is "error", write to error table and info table. but log is "info", write to only info table. is it right?
Hi, I'm curious that in 'error' log, why is it showing stack trace only for the last error log only (where we have created Error object new Error? but not for the above error (logger.error('text error') ?
Hi, if you execute logger.error('error message') instructs the logger to just write the string to the log file with the error severity. If you have an error object, you also get context, i-e. the stack trace
Hi Supun, you could check out OpenTelemetry for this and then add the trace id in a custom log format in winston: github.com/open-telemetry/opentelemetry-js
The examples show the statements logger info, logger.warn etc... and prints them with meta information. How would you use this with actual code? I'm trying to figure out how to have the console print the specific error when a Promise is rejected or when it otherwise fails.
Hi Jonathan, if your promise gets rejected you probably want to log error you got back. Maybe this Stackoverflow question helps: stackoverflow.com/questions/56090851/winston-logging-object You can use string interpolation for the error message and you could add the error as a second parameter. Like so the entire object gets attached to the metadata of the log statement.
Hi Dimuth, the simplest way would indeed be to just use one logger instance. Theoretically - if you wanted to - you could also create a logger factory that creates loggers. Like so you can add additional meta info (e.g. the name / path of the file from where the log command was invoked so you find the location where the log was produced more easily
This video was just perfect for logging in production, thank you! I just encountered a small issue. When production environment, the error stack will appear in console but it won't be saved into combined.log, any ideas? My code is pretty much the same, simply adding one more transport layer to the transport array. ``` transports.File({ filename: 'combined.log', level: 'error' }) ``` Any ideas why it doesn't write the err stack into the file but it does write everything else? Thank you!
What do you think about this video?
Please let me know in the comments below.
After so many try o google search, I found your video on logging in node js. Nobody explain this stuff so simply. Thanks!
You are most welcome!
Great to see someone actually explaining what is going on, rather than just walking through the code.
Hands off the best winston tutorial I've seen so far. Beautifully simple but advanced insights on production setup.
Attaboy
Glad you enjoyed it!
Great stuff. I'm currently trying to build a REST API node backend, and have been frustrated with most tutorials giving overly basic examples that don't help with real-world application.
This was comprehensive, crystal clear, and without fluff. You've probably saved me a week of banging my head on my keyboard, cheers!
Your videos are really great; very easy to follow and you take time to explain things and not just quickly gloss over them. Thank you!
great, thx
For a junior dev with no real mentor, this video was a godsent. Thank you!! 🙏
thx, you're welcome Nawaz 👍
really comprehensive
please produce more content like this
thx
I was just about to setup Winston for a project and I saw this, great video!
thx Godson, I'm glad it helped you to save some time 👍
I'm working on implementing Winston for a project, and found their "Quick Start" code confusing. Thank you so much for creating this video tutorial and breaking down what's happening even further. This was incredibly helpful!
This was helpful for someone who attempted to log their stuff for the first time. Though there could be 30 more minutes on when to log, what to log, how to log properly etc.
Hi! Is it possible concatenate some custom log in the end of main log for each route?
Than for this awesome tutorial on Winston. It was very precise and contained everything I needed to setup a logger in local and production
thx Prabhu 👍
Thx I'm about to use winston as my logger on a project.
Maybe for the future also discus setting up the daily rotation
Thank you very much, clean and clear. Can you please help how to set winston globally in express node js application?
Thank you for amazing video!,but I was wondering how am I going to log the action which happen in different routes ?
Great timing, just had to built a new node application, but without express. For express usually morgan is my logging tool.
nice 👍
Great Video for getting to know Winston and some first best practices. Thanks!
you're welcome 👍
Great video! It helped me with a task I had at my job
Keep up the great work! :)
thx Leonard 👍
Noticed at 26:20 you needed to restart the app because you're making changes on the fly, is there a way to apply dynamic logging without restarting the app ??
So cool! Watching you from Ukraine. I love that you allways show how it has to be in real project and showing package structure. It really helps me on my first job!
Glad you enjoyed it!
Amazing intro on Logging!!!
thx Suyash 👍
How to daily rotate logs using Winston except the first day?? Like my initial log should be info.log then next day a new info.log should be created and previous one should be appended with a date
per default, the log is written to stdout so that it is just getting appended to the existing logs. This is mostly sufficient because you have everything in one place. Not sure if there is a use case for what you want to do because each log statement also contains a timestamp
Got me going with Winston really fast. Good job
thx, I'm glad it helped 👍
Great video, thanjs so much. Would Winston work with Heroku deployed apps? or maybe there are some adittional setting to do. Thanks!!
sure. As long as the service you are running it on can run Node, Winston, as well as any other library is fine.
Great stuff, very useful to get up and running without needing to spend additional time in docs!
Glad it was helpful!
Thank you for making this.
you're most welcome
Very well explained
thx Tanishq, I'm glad it was understandable 👍
great tutorial! many thanks!
Glad you enjoyed it!
Great video! Thank you!
Glad you liked it!
Great video, thanks!
Thx Mikhael, I'm glad you liked it!
really great tutorial
Glad you liked it
real good and concise video! thanks
Glad it was helpful!
This video has been very helpful.
thx I'm glad it was helpful 👍
@productioncoder How about storing logs in Azure blob or any table ?
Azure offers a dedicated log aggregator for this. Typically you do not need to configure anything, but just run a daemon in your cluster. The daemon will grab all output from std out and push it to the log aggregator
Such a great video. Been watching a lot of your videos and hopefully I'll be coding on prod soon :)
thx, I'm glad it helped
Great video! I use to use loggrus in golang, but now I'm rolling on Nodejs and express and this it's very similar.
thx Abel 👍 Yeah loggrus is a great Go logging library
Awesome video, Kindly also make production logging with winston and elk, with also tracing the whole api request cycle, I have been looking for child loggers and tracing whole api request ?
Hey, thanks. Have you worked on creating custom transports in Winston?
not really, using custom transports so far was no needed, almost all standard use cases work out of the box
Great video as always, could you do one for the "morgan" package as well?
hi Osama, I will add this to my list of potential topics 👍
How do you add linenumber and function name in logger ?
Thank you for this great video!
you're welcome JCube 👍
It would be great if you can show to save the log in the database.
nice video and good explanation
thx SaiRam 👍
tks -> is it possible to get the line number
Great video. You rock!
thx Gustavo 👍
maja aa gaya bhai thank you
It would have been the best tutorial video but the light themed github got me
😁
Thank you, I learned a lot 👍
Glad it was helpful!
Please make a video on logging in react. There is no video about it on internet.
Its was a really helpful video.. just wanted to know how to configure splunk cloud with it?
Hi Anandita, I haven't use Splunk Cloud so far
Hi! Thanks for the great video!
I am using nestjs as my backend framework and I wonder how I can add a trace id to each incoming request?
Hi Shimron, assigning it a traceId would do the trick. This works especially well with a microservice-based system. Maybe check out opentelemetry.io/. If you only have 1 server / microservice, opentelemetry might be overkill though. It is rather meant for microservices
Amazing Explanatiion Thanks
thx cuberos, I'm glad it was helpful!
Wow, such an easy and wonderful video. Thanks for it. :)
Can we make the logger work for any specific level, like just for error or warn, excluding the other higher levels? Please suggest something.
Hi Anshika, yes you can do that. If you only want to output warning or anything more severe, then you would need to set level: warning when you create the logger.
So basically this level field means that it will log for warn logging level and below. I.e. you would log for warn and error. You can find a more detailed description here: github.com/winstonjs/winston#logging-levels
Hi,Do you have any idea how to log every operation happened in the database and save it into the MySql Database?
I've noticed when using with morgan that the morgan console log appears after winston. I'm assuming it's because morgan is async and winston is not? Do you know a good way to adjust winston logs to appear after morgan request logs? I think it looks nicer when you see the request and then all the details that follow in your logs.
Hi Nathan, I think it depends when the middleware is plugged in. If you plug it in right at the start then morgan logs should be the first things that show up: github.com/expressjs/morgan#expressconnect
Good job! And thank you very much!
you're most welcome 👍
amazing! thanks a lot
Glad you liked it!
Thanks for the Video. Quite helpful . I had a doubt, Should the logs be stored in a log file or is it better to use mongodb for storing ?
Hi Arun, storing it in a file is probably the cheaper option. If you don't intend to search in the logs dynamically from some webapp, then I guess you can just run go with a file. Storing the logs in MongoDB (in a separate instance from your main db so it does not affect performance) would only make sense if you have a use case where you need to dynamically access the logs from your webapp and search in them and so on
@@jgoebel Thank you for responding and clarifying my doubt.
@@arunkutz you're most welcome Arun 👍
i guess,
logger.log('error',`error string`);
logger.error('error string');
those commands do samething.
if log is "error", write to error table and info table.
but log is "info", write to only info table. is it right?
if your log level is error, then you would only see the error logs. If your log level is info, you would see both the error and the info logs
@@jgoebel thank you
Hi, I'm curious that in 'error' log, why is it showing stack trace only for the last error log only (where we have created Error object new Error? but not for the above error (logger.error('text error') ?
Hi, if you execute logger.error('error message') instructs the logger to just write the string to the log file with the error severity. If you have an error object, you also get context, i-e. the stack trace
Nice tutorial. But I was looking for outputting the log data into a file. Could you please explain that part.. Thanks
Hi wisement, you can add a filetransport if you want to do this: github.com/winstonjs/winston#creating-your-own-logger
@@jgoebel , thank you. I did that and it worked..
I want to track all logs in one HTTP request flow and for that, I want to add Correlation Id to the log. is there have any way to do this with Winston
Hi Supun, you could check out OpenTelemetry for this and then add the trace id in a custom log format in winston: github.com/open-telemetry/opentelemetry-js
Ok fine! Take your 47'th like! You deserve it ;)
Seriously now, thank you bro. One hell of an explanation.
haha thx Tal, I'm glad it was understandable 👍
Amazing, ty :)
thx
The examples show the statements logger info, logger.warn etc... and prints them with meta information. How would you use this with actual code? I'm trying to figure out how to have the console print the specific error when a Promise is rejected or when it otherwise fails.
Hi Jonathan, if your promise gets rejected you probably want to log error you got back. Maybe this Stackoverflow question helps: stackoverflow.com/questions/56090851/winston-logging-object
You can use string interpolation for the error message and you could add the error as a second parameter. Like so the entire object gets attached to the metadata of the log statement.
When I use winston all over the files, should I create a singleton object for this or something like that?
Hi Dimuth, the simplest way would indeed be to just use one logger instance. Theoretically - if you wanted to - you could also create a logger factory that creates loggers. Like so you can add additional meta info (e.g. the name / path of the file from where the log command was invoked so you find the location where the log was produced more easily
This video was just perfect for logging in production, thank you! I just encountered a small issue. When production environment, the error stack will appear in console but it won't be saved into combined.log, any ideas? My code is pretty much the same, simply adding one more transport layer to the transport array.
```
transports.File({
filename: 'combined.log',
level: 'error'
})
```
Any ideas why it doesn't write the err stack into the file but it does write everything else? Thank you!
Same. transports.File isn't working for me too. Were you able to resolve this?
@@yetanothercodingchannel2287 Unfortunately, no. I had to give up on this.
Greate video 😍
thx Morteza
Hey, how do we save the log file with the timestamp?
You can specify the filename when you create a file transport: github.com/winstonjs/winston#usage
Make a more detailed video on it. Like how to make a more production-ready logger with tools like Datadog or Splunk
thx Suyash for the idea, will add it to my list of potential topics 👍
You are look like a Elon Musk Brother
I'm still working on my first rocket though 😁
An excellent video, thanks.
Glad it helped